251956 - crash after login information is entered into plesk

Status: RESOLVED DUPLICATE of bug 263603

(SeaMonkey :: General, defect)

Description

[tom]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a2) Gecko/20040714
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a2) Gecko/20040714

After logging into Plesk, mozilla crashes with this error, The instruction at
"0x77f596cf" referenced memory at "0x00000001". The memory could not be "written".

http://www.sw-soft.com/en/products/plesk7/demo/

at the above url, you can test with a demo of plesk. 

Reproducible: Always
Steps to Reproduce:
1. log into plesk
2.
3.

Actual Results:  
mozilla crashes with this error - The instruction at "0x77f596cf" referenced
memory at "0x00000001". The memory could not be "written".

Expected Results:  
gone to the home page of the plesk control panel.

Comment 1

[Adam Hauner]

Tom: Could you provide TalkBack incident ID?

Comment 2

[Arthur]

Login at http://plesk7r.demo.sw-soft.com:8443/ with "admin" and password "plesk".

Confirming on Linux 2004112506 (gtk2 and gtk1 builds). Only crashes on gtk2
build, freezes the gtk1 build. No talkback reports are generated. If the bug
doesn't get triggered on first try, hit reload once or twice.

Comment 3

[Arthur]

Stack and QA wanted. Can't confirm because I don't know what component to move
to, "General" is missing for Core, which is certainly appropriate for product as
it is reproducable on Firefox trunk-nightly 20041125.

Comment 4

[Arthur]

CC'ing the owner of the Firefox general component and asking whether this blocks
1.7.5 (should in my opinion as plesk gets used by a lot of hosting companies,
which means that a lot of website owners would see this bug).

Comment 5

[timeless]

current rule (as of some time yesterday) was that 1.7.5 was supposed to be crash
equivalent to ff1.0. if you crashed ff1.0, then we were not to make you not
crash in 1.7.5.

Comment 6

[Arthur]

I would be very much in favor of making it crash equivalent to ff1.0 as it only
crashes on the latest trunk nightlies but not on ff1.0. ;)

Comment 7

[Arthur]

Removing request for blocking 1.7.5. 1.7.x branch build 20041126 doesn't show
this bug. Only on trunk. Adding blocking 1.8a6 request and regression keyword.

Comment 8

[Andrew Schultz]

I got this from valgrind:

==22957== Invalid write of size 4
==22957==    at 0x1BA7F19C: js_Interpret (jsinterp.c:4422)
==22957==    by 0x1BA71FAA: js_Execute (jsinterp.c:1516)
==22957==    by 0x1BA4D389: JS_EvaluateUCScriptForPrincipals (jsapi.c:3702)
==22957==    by 0x1CE58F82: nsJSContext::EvaluateString(nsAString const&, void*,
nsIPrincipal*, char const*, unsigned, char const*, nsAString*, int*)
(nsJSEnvironment.cpp:988)
==22957==  Address 0x1ECF8D40 is 11496 bytes inside a block of size 11499 alloc'd
==22957==    at 0x1B904A80: malloc (in /usr/lib/valgrind/vgpreload_memcheck.so)
==22957==    by 0x1BA4E0A6: JS_ArenaAllocate (jsarena.c:194)
==22957==    by 0x1BA70260: js_AllocRawStack (jsinterp.c:334)
==22957==    by 0x1BA71E50: js_Execute (jsinterp.c:1468)

indicating bug 263603.  I verified that loading

http://plesk7r.demo.sw-soft.com:8443/javascript/conhelp.js.php

in JS Shell under valgrind triggers the error mentioned in bug 263603 comment 9

*** This bug has been marked as a duplicate of 263603 ***